summaryrefslogtreecommitdiff
path: root/src/python/multifasta_to_otu.py
diff options
context:
space:
mode:
authorCalvin <calvin@EESI>2013-03-12 14:14:59 -0400
committerCalvin <calvin@EESI>2013-03-12 14:14:59 -0400
commita200b2df4212521b7e006f17979d7c4c2d4b8ca9 (patch)
tree0d3eb315ddac74cb710020da674b42236b315a5c /src/python/multifasta_to_otu.py
parent847cb4aa7f683254118605fd7f47d0251519e1ea (diff)
removed multifasta's py extension, modified the --help argument output for quikr_train
Diffstat (limited to 'src/python/multifasta_to_otu.py')
-rwxr-xr-xsrc/python/multifasta_to_otu.py154
1 files changed, 0 insertions, 154 deletions
diff --git a/src/python/multifasta_to_otu.py b/src/python/multifasta_to_otu.py
deleted file mode 100755
index c6fb562..0000000
--- a/src/python/multifasta_to_otu.py
+++ /dev/null
@@ -1,154 +0,0 @@
-#!/usr/bin/python
-
-from multiprocessing import Pool
-from Bio import SeqIO
-import multiprocessing
-from subprocess import *
-import os
-import gzip
-import quikr as q
-import sys
-import numpy as np
-import argparse
-import platform
-import csv
-
-# our defaults
-kmer = 6
-lamb = 10000
-output_directory = ""
-input_directory = ""
-
-
-def main():
-
- global kmer
- global input_directory
- global output_directory
- global lamb
- global trained_matrix
-
- parser = argparse.ArgumentParser(description="MultifastaOTU")
-
- parser.add_argument("-i", "--input-directory", help="directory containing fasta files", required=True)
- parser.add_argument("-o", "--otu-table", help="otu_table", required=True)
- parser.add_argument("-t", "--trained-matrix", help="your trained matrix ", required=True)
- parser.add_argument("-f", "--trained-fasta", help="the fasta file used to train your matrix", required=True)
- parser.add_argument("-d", "--output-directory", help="quikr output directory", required=True)
- parser.add_argument("-l", "--lamb", type=int, help="the default lambda value is 10,000")
- parser.add_argument("-k", "--kmer", type=int, help="specifies which kmer to use, default=6")
- parser.add_argument("-j", "--jobs", type=int, help="specifies how many jobs to run at once, default=number of CPUs")
- args = parser.parse_args()
-
- # our defaults
- jobs=multiprocessing.cpu_count()
- trained_matrix = args.trained_matrix
- input_directory = args.input_directory
- output_directory = args.output_directory
-
- # Make sure our input exist
- if not os.path.isdir(args.input_directory):
- parser.error("Input directory not found")
-
- if not os.path.isdir(args.output_directory):
- print "Output directory not found, creating directory"
- os.path.mkdir(args, output_directory)
-
- if not os.path.isfile(args.trained_matrix):
- parser.error("Custom trained matrix not found")
-
- if not os.path.isfile(args.trained_fasta):
- parser.error("Fasta file of trained matrix not found")
-
- # use alternative lambda
- if args.lamb is not None:
- lamb = args.lamb
-
- if args.jobs is not None:
- jobs = args.jobs
-
- if args.kmer is not None:
- kmer = args.kmer
-
- # Load trained matrix
- if qu.isCompressed(args.trained_matrix):
- trained_matrix_file = gzip.open(args.trained_matrix, "rb")
- else:
- trained_matrix_file = open(args.trained_matrix, "rb")
-
- trained_matrix = np.load(trained_matrix_file)
-
- # Return a list of the input directory
- fasta_list = os.listdir(args.input_directory)
-
- # Queue up and run our quikr functions.
- pool = Pool(processes=jobs)
- results = pool.map(quikr_call, fasta_list)
-
- # Create an array of headers
- headers = []
- trained_matrix_headers = open(args.trained_fasta, "rU")
- for header in SeqIO.parse(trained_matrix_headers, "fasta"):
- headers.append(header.id)
- trained_matrix_headers.close()
-
- # create our number of reads matrix
- number_of_reads = np.zeros((len(headers), len(fasta_list)))
-
- # load the keys with values from each fasta result
- for fasta, fasta_it in map(None, fasta_list, range(len(fasta_list))):
-
- count_sequences = Popen(["grep", "-c" , "^>", args.input_directory + fasta], stdout=PIPE)
- number_of_sequences = int(count_sequences.stdout.readline())
-
- proportions = np.loadtxt(output_directory + fasta);
-
- for proportion, proportion_it in map(None, proportions, range(len(proportions))):
- number_of_reads[proportion_it, fasta_it] = round(proportion * number_of_sequences)
-
- # remove empty rows from our matrix
- final_headers = list()
- final_data = list()
- for row, header in map(None, number_of_reads, headers):
- if row.sum() != 0:
- final_headers.append(header)
- final_data.append(row)
-
- # convert from a list back into a numpy array
- final_data = np.array(final_data, dtype=int)
-
- # stack our final header and our number_of_reads matrix
- number_of_reads = np.column_stack((final_headers, final_data))
-
- # write our OTU table
- output_file = open(args.otu_table, "wb")
- writer = csv.writer(output_file, delimiter="\t")
-
- #write out our fasta file row
- writer.writerow(['# QIIME vGail OTU table'])
-
- fasta_row = ['#OTU_ID']
- fasta_row.append(' '.join(fasta_list))
- fasta_row = [' '.join(fasta_row)]
- writer.writerow(fasta_row)
-
- # write out our results
- for i in range(0, np.shape(number_of_reads)[0]):
- writer.writerow(list(number_of_reads[i]))
-
- output_file.close()
-
- return 0
-
-
-def quikr_call(fasta_file):
- input_location = input_directory + fasta_file
- output_location = output_directory + os.path.basename(fasta_file)
-
- xstar = q.quikr(input_location, trained_matrix, kmer, lamb)
- np.savetxt(output_location, xstar, delimiter=",", fmt="%f")
- return xstar
-
-if __name__ == "__main__":
- sys.exit(main())
-